[comp.protocols.appletalk] NCSA Telnet, KIP, and subnets

brad@saturn.ucsc.edu (Brad Smith) (10/27/87)

Does anyone know how to get NCSA Telnet to access a subnet of
a class B net (i.e. subnet mask 255.255.255.0)?  I have tried
all kinds of values in the "Use Subnetting Mask" box of NCSA
Telnet version 2.0 (a great piece of software at the perfect
price!), including "ffffff00", "ff.ff.ff.0", "0xff.0xff.0xff.0",
all of the previous with "F"s, and 255.255.255.0.  All cases
seem to have no effect; from what I can see (with tcpdump on
a Sun) it looks like the Kbox catches the packet, makes its own
decision as to whether it is going to a different network (which
doesn't detect subnets) and arps for the host on the attached
Ethernet (which the host isn't attached to).

The Kbox is running release # 5 of the KIP code (I assume...
the srec file I load to it is "at-gw.srec.5").

Any help will be greatly appreciated,

Brad Smith
System Administrator
CIS Board, UC Santa Cruz
brad@saturn.ucsc.edu

hpoppe@scdpyr.UUCP (Herb Poppe) (10/28/87)

In article <1032@saturn.ucsc.edu>, brad@saturn.ucsc.edu (Brad Smith) writes:
> Does anyone know how to get NCSA Telnet to access a subnet of
> a class B net (i.e. subnet mask 255.255.255.0)?  ...

Try the hex value "08" (zero-eight). I guess this means "eight bits on".
Not my idea of a mask either.  It took me awhile to stumble on this one!
The documentation needs to be more explicit on the appropriate value
to enter for this parameter. Are you listening, Gaige?
-- 
Herb Poppe      NCAR                         INTERNET: hpoppe@scdpyr.UCAR.EDU
(303) 497-1296  P.O. Box 3000                   CSNET: hpoppe@ncar.CSNET
		Boulder, CO  80307               UUCP: hpoppe@scdpyr.UUCP

rick@sbcs.UUCP (10/30/87)

While we're on the subject, does anyone know if the sources for NCSA telnet 
are available?

						Rick Spanbauer
						SUNY/Stony Brook

gaige@ncsa.uiuc.EDU.UUCP (10/30/87)

The current format for NCSA Telnet subnet masks is eight hexidecimal digits
indicating an actual bit mask for the IP number.  The most common example,
for a class B net with eight bits of subnetting is "ffffff00", which happens
to be the same as class A, sixteen bits of subnet.  Similarly, class A with
eight bits is "ffff0000."

It is important to note that if the kinetics box stradles subnets (has one
IP net number on the ethernet side and a different one on the AppleTalk side),
the AppleTalk side of the Kinetics box must be specified as a gateway in the
config.tel file.  An example of this is a KBox running the combined gateway from
kinetics with 128.174.20.1 Ethernet and 128.174.22.1 AppleTalk.  Assuming a
class B, eight bit subnet, the subnet mask would be "ffffff00" and the following
entry should be in the config.tel:

name=Kbox  hostip=128.174.22.1 gateway=1

I hope this helps,
Gaige B. Paulsen
National Center for Supercomputing Applications
University of Illinois

gaige@ncsa.uiuc.edu